The objectives of this study were to evaluate the production, the land use efficiency, the financial return and the risk of investment in agroforestry systems (AFS) traditionally used in the State of Rondônia. Data of the agroforestry systems were from
a 15-year agroforestry essay, owned by Embrapa, installed in the Experimental Field of Machadinho d’Oeste, in the Northeast of Rondônia State. The experimental design was in randomized blocks with four replications and the analysis followed the split-plot
scheme, being the plot made up of eight production systems and five monocrops, and in the subplot the time of production was studied. The species used were: banana (Musa sp.) - Ba, black pepper (Piper nigrum L.) - Pm, cupuaçu (Theobroma grandiflorum (Willd. ex Spreng.) Schum.) - Cp, Brazil nut (Bertholletia excelsa H.B.K.) - Ca, freijó (Cordia alliodora ((Ruiz & Pav.) Oken) - Fr and pupunha (Bactris gasipaes Kunth) - Pu. The treatments were the following agroforestry systems: T1 Ca-Ba-Pm-Cp, T2 Fr-Ba-Pm-Cp, T3 Pu-Ba-Pm-Cp; and monocrops: T4 Ca, T5 Fr, T6 Pu, T7 Ba and T8 Pm. Agroforestry systems (AFS) data were analyzed in a comparative way with those from monocrops, where the situation observed for the AFS was the desirable one. The land efficiency with agroforestry systems in relation to monocrops was evaluated through the Relative Productivity (RP) and the Land Equivalent Ratio (LER). The financial analysis of the agroforestry systems were supported by the economic criteria for forest projects evaluation. Data were analyzed from the cash flow formation, including the costs and incomes along the horizon of planning of each production system, where the situation of higher economic return was considered the desirable one. The risk of investment was made through simulations with the software @RISK for the agroforestry system with the main financial results. The main conclusions were that the tested species are appropriated for association in multiestrata agroforestry systems; the stand density presented a factor with great influence on the productive behavior of the agronomic species; the association of species did not affect the growth of Brazil nut (Bertholletia excelsa) and freijó (Cordia alliodora); being necessary to keep on monitoring the production of Brazil nut fruits for better conclusions on its productive behavior; the results of RP and LER showed that SAF T1 Ca-Ba-Pm-Cp is the best production alternative, in relation to SAFs T2 Fr-Ba-Pm-Cp e T3 Pu-Ba-Pm-Cp. It was also concluded that agroforestry systems were, along a continuum 10-year period, a land use system more efficient than monocrops; the management and harvesting costs represented more than 70% of the total costs composition; the labor costs participation was higher than 50% in the site preparation and long-term maintenance phases of the agroforestry systems; all the production systems were considered economically viable, being T1 Ca-Ba-Pm-Cp the agroforestry system with the best financial results, and, even though the high costs, the investment risk in this agroforestry system was effectively small, according to the risk simulation and the results of the financial criteria applied.
